Institutions,Ideology,and Nonideal Social Ontology

1. Idealization and the Problem of Ideology

The social world is a messy place. While most of us are still able to navigate it most of the time, few of us would be able to give a detailed account of its fundamental structure or its basic constituents. This is where social ontology potentially comes in. While the social sciences tend to deal with concrete social phenomena, often in detailed and highly specific ways, social ontology can be understood as the most abstract and general form of social theory. This also means that it is difficult to see how social ontology can do without at least some degree of idealization. ² We might differ on how far we want to push such idealizations, but the goal is presumably to sharpen the contours of social reality, allowing us to see what is common to a variety of social and institutional arrangements, and any such commonalities are bound to lie on a high level of abstraction.

This need for idealization, but also an idea about this not being inherently problematic, has been laid out by Margaret Gilbert (2006, 180), in developing her plural-subject approach:

The plural subject account of a society provides an articulated model in relation to which the realities of the world can be assessed. The perfect case is, if you like, an ideal type in the sense of Max Weber. It is not necessarily a good thing—it is not ideal in that sense. Rather, it is the clearest possible case of its kind.

And yet, while Gilbert is arguably correct in understanding ideal types as not normative in the sense of implying recommendations, are matters really this straightforward? While Weberian ideal types are not strongly normative in the sense of being prescriptive, they can still be understood as weakly normative in the sense that they create a hierarchy where, in this case, certain conventions, institutions, or even societies come to be classified as bona fide examples, others as relatively normal, and further others as exceptions or anomalies. An ideal type is an interpretative tool which we use to simplify something more complex to understand it better, and since such complex cases will often be open to different interpretations, the choice of ideal type will have implications for how we read many cases, such as which cases that come out as typical or atypical.

This observation potentially turns into a problem because some real-life political disputes tend to partly hinge on whether certain phenomena are endemic or marginal, or whether they even count as institutional matters at all. ³ For instance, for the question about how serious race and gender injustices are in contemporary societies, it can be highly important whether racism and sexism are pervasive phenomena or primarily something characterizing a few bad eggs. How we understand what social relations and institutions are like typically will have an impact on how it makes sense to answer that kind of question. At least when it comes to the understanding of society that underpins much of liberal political theory, Charles W. Mills (2005, 168) raises this kind of worry:

An idealized social ontology of the modern type . . . will typically assume the abstract and undifferentiated equal atomic individuals of classical liberalism. Thus it will abstract away from relations of structural domination, exploitation, coercion, and oppression, which in reality, of course, will profoundly shape the ontology of those same individuals.

Mills goes on to argue that the kind of basically egalitarian framing of social relations which characterizes the standard models underpins a framing of phenomena like racism and sexism as anomalous, and there is accordingly a risk that our “opting for ‘ideal’ theory has served to rationalize the status quo” (Mills 2005, 181). While Mills is not discussing social ontology in terms of the distinct field that goes by that name, the relevant idealizations can clearly be found there as well.

If we look at analytic social ontology, its typical explanatory project is arguably to understand how a set of distinct individuals can come to join into some form of “we” or to share some idea, like certain rules, in a way that goes beyond merely having the same idea. Notable theorists like Gilbert (1989, 2006), Tuomela (2002, 2013), and Searle (1995, 2010), possibly the three most important figures in analytic social ontology, can all be said to represent this orientation, albeit in different ways. ⁴ Another, although partly overlapping, the main strand in analytic social ontology takes a game-theoretical approach, focusing on instrumental rationality and equilibrium solutions as a way of understanding why certain rules can spontaneously be established as the ones that regulate the relations of a set of individuals. Lewis (1969) was an early representative, but in more recent years there has been renewed interest in this kind of modeling, as exemplified by Guala and Hindriks (2015) and Guala (2016), as well as Smit, Buekens, and du Plessis (2011, 2014). ⁵

While there is no room here to go into the details of these accounts, it would seem that in being oriented toward showing how people without institutions or people not already forming a social group can ex nihilo establish an institution or form a social group, the underlying logic of institutions that is primarily sought to be explicated is a logic of creation. In approaching the explanatory task of social ontology in this way, one might discern a similarity with the contractarian tradition in political theory, where theorists have attempted to explain the legitimacy of government through how people without government would prefer to have government in place or agree to subject themselves to political authority. Contract theorists do not take our actual starting-point, individuals already living in an established society, as their starting-point, although they can be understood as taking a more difficult starting-point by starting from scratch rather than helping themselves to preexisting institutions. ⁶ The underlying methodological idea is that if you can show how a political union can or would arise even under conditions which are idealized in the sense that there is no prior social basis for that union, you have also thrown light on actual political unions (and made a general case for the legitimacy of government authority).

Similarly, by abstracting from existing social relations and showing how otherwise atomic actors can unify or form a relevant whole, a social ontology focused on the logic of creation can throw light on our actual social groups or institutions. And, also in parallel to contract theory, by answering this kind of how-possibly question, ⁷ the models that analytic social ontologists tend to provide can furthermore be seen as explaining the normativity of institutions: if we can show how atomic actors can become bound to each other in certain ways, then that can be seen as throwing light on how actual people are bound. If the underlying logical structure of our institutions is explicated in terms of how they have been collectively instituted, how we are jointly committed to certain rules regulating our affairs, or how these rules represent a mutually beneficial solution to a problem of coordination, individuals that participate in such institutions can, in turn, be understood as having reasons for playing by the rules. Institutions also appear to be characterized by a relative stability and a capacity for self-reproduction, and the idea that they represent agreements or equilibrium solutions can help explain this too. However, even if the standard approach in analytic social ontology has these potential explanatory advantages, it would also seem that to the extent that we think that these particular how-possibly explanations throw light on our actual institutions, we are also establishing a situation where these institutions are cast in a prima facie rosy light simply qua being institutions. And the reason is that since the model has assumed individuals to be not already constrained, and hence at liberty to either take on such constraints or not, this means that for them to be reasonably conceived of as choosing to take on certain constraints, there must be some gain involved, or at the very least some form of consent.

Raimo Tuomela (2013, 229), for instance, has made this quite explicit, suggesting that institutions tend to solve or dissolve “conflicts between collective and individual rationality—and give cooperative, collectively beneficial solutions to these problems in the face of the chaos and conflict that unfettered individual action tends to lead to.” There is nothing puzzling about this statement given that the explanatory task has been understood in terms of showing how previously unrelated or uncoordinated individuals can come to form a collective or create certain institutions, because then the relevant comparative contrast for making this value judgment becomes not having any institutions at all, and at least on an abstract level institutions accordingly come out as typically benign, either by being rooted in some form of agreement or in some form of benefit (or both). This way of understanding the explanatory task also means that to the extent that institutions involve power relations, power will tend to come out as the product of certain institutions having come into place rather than as an antecedent factor shaping those institutions. ⁸

To be clear, relying on this type of approach certainly does not commit anyone to the idea that institutions are always beneficial or based on choices that individuals have made together in some substantial sense. But it does mean that one has adopted an approach according to which certain kinds of problematic institutions become conceptualized as marginal or even deviant cases, ones where the core logic of institutions have been twisted out of shape, perhaps not even counting as being institutional matters. And this is the worry raised by Mills. In using ideal types in understanding our own societies, we sharpen the contours of the social reality we inhabit, but even setting aside the fact that using ideal types always involves sorting cases into typical and atypical, the suggestion here has been that the specific character of the ideal types often employed in analytic social ontology potentially has the consequence that we will sharpen these contours in a way that means that the default interpretation of institutions, including those that we ourselves inhabit and occupy specific positions in, is a pretty-picture interpretation.

One possible response here is that if institutions really do typically have a beneficial character, then even if relying on a certain ideal type can in some cases obfuscate the workings of some institutions, it could on the whole still be a reasonable understanding of institutions. But apart from the problem that an assessment of an institution as beneficial presupposes a choice of relevant comparisons (and also invites questions about for whom they are beneficial), it seems clear, especially if we look historically, that strongly hierarchical social relations and repressive institutions are hardly anomalies. For instance, bonded labor has been common practice across the world; in many places, we have even had chattel slavery, and serfdom was a mainstay of feudal Europe. There have been (and still are) systems of forced marriages, even marriages by abduction, honor killings, and female genital mutilation. These are all institutions. And while social ontologists tend to focus on examples such as the presidency of the United States, history has been filled with emperors, kings, princes, dictators, and chieftains who have rested their power more on the sword than on agreements or mutual benefit. And yet these are often also offices within well-established institutions.

To the extent that such repressive institutions have been relatively stable, it does seem reasonable to assume that there has been a certain level of acceptance of them, but there really seems to be no need to assume anything more than widespread compliance to account for this. And while there is a tendency in analytic social ontology to conceptualize participants as more-or-less equals, or at least as playing the same games of coordination, these historical cases are examples of thoroughly stratified societies, characterized by divisions rather than unity—where force is clearly a more important factor than consent. This is not to say that the reasons why people comply cannot be understood in terms of the incentives they have, but their incentives would seem to be strongly shaped by the institutions in place, and acknowledging this is a very different thing from saying that the institutions are grounded in how they solve a game of coordination understood in terms of some set of preinstitutional incentives. While it would arguably be correct to say that many oppressive historical institutions have had as an important function to coordinate farming practices, and where food production is probably as close as we can come to a preinstitutional incentive, this would still only amount to an explanation of the fact that there are some institutions in place, but where there could be several different institutions which would fulfill the relevant function. To then model institutions as typically involving some sort of collective agreement or as involving implicit joint attitudes or as amounting to an equilibrium solution in a game-theoretical sense will not obviously always be the best interpretation.

Of course, if we limit ourselves to those contemporary liberal democracies where social ontologists tend to live, these models might have a better fit with institutions there, but apart from us not wanting simply an account of institutions in liberal democracies, this move would be unlikely to persuade someone like Mills. The worry about social ontology playing the role of ideology is, after all, mainly a worry about how it contributes to painting a misleadingly pretty picture of institutions in those very societies. If we take this worry seriously, there can probably be a number of different ways to address it within those more precise accounts which have been lumped together in the above, and the argument here has been far too broad to establish anything more than that there is a need to tread carefully in this respect when constructing an account of institutions. However, to better understand our options, it would be valuable to also look at other types of accounts, and the following two sections will move toward one such possible type of account, one that still clearly falls within the bounds of analytic social ontology, but which hopefully steers clear of potentially playing an ideological role.

2. Toward a Nonideal Social Ontology

In discussing how political theory should deal with the problem of ideology, Mills suggests that a move toward nonideal theorizing is reasonable. Of course, it would be difficult to do social ontology without any idealizations, but in the present context Mills’s suggestion can arguably be understood as pointing to the need for a type of social ontology where we understand institutions in a way that leaves no interpretative slant in terms of whether institutions are typically or atypically benign/beneficial or harmful/oppressive. There would seem to be at least two steps that should then be taken.

To begin with, as a methodological principle, one needs to be careful about the diet of examples used in doing social ontology. There has arguably been a tendency in social ontology to rely on fair-weather examples, like taking walks or painting houses together, in constructing the basic elements of the relevant models. But the problem with these examples is that they often involve people who are more or less equals doings things in robustly consensual ways. At the very least, that kind of diet of examples seems to provide no safeguard against producing theories that idealize institutions in the sense that institutions come out as typically benign or even beneficial. To do nonideal social ontology, in contrast, will instead involve relying heavily on examples about unequal social relations, on testing models by explicating situations involving dominance and oppression, where some people are privileged and others subordinated. Of course, this is not to say that these are the only examples that matter, the point is not to replace one one-sided diet of examples with another, but if we want to avoid ending up in accounts that idealize institutions, an important corrective would seem to be to maintain an emphasis on these messy examples.

In addition, and more substantially, one would shift one’s understanding of human beings from one that pictures them as fundamentally atomic, abstracted from their social relations, to one that pictures them as fundamentally social. The explanatory task that we face is not a mystery of the social, where we need to understand how previously unrelated individuals can form social groups or construct institutions. We would start instead from an understanding of human beings as always already social beings, always already entangled in a host of social relations, and always already positioned within institutional frameworks that are more or less set. What always already social means is not simply that that we are social in an abstract way, but that we are always already socialized into certain concrete modes of relating to other people, and where these modes will be part of how we understand the situations we face and which actions that present themselves as meaningful alternatives to us. This also means that how we understand problems of coordination that we face will always already be infused with our understanding of how we typically (and concretely) coordinate. ⁹ And with respect to institutions, actual individuals find themselves already brought up within certain institutions or having moved to a society with already existing institutions firmly in place. This does not mean we cannot identify any logic to institutions, but that in doing nonideal social ontology, we would not understand institutions through explicating a logic of creation.

Moving toward this kind of nonideal social ontology does not mean abandoning the idea that institutions ultimately rest on some form of recognition among participants in the activities that such institutions regulate or even partly constitute. We seem to be able to identify at least two core elements to institutions: (a) some specific type(s) of objects of recognition, and (b) some specific mode(s) of recognition. The parts of the standard models that were found to be problematic in the above were primarily their attempts at conceptualizing the relevant mode(s) of recognition as more or less some form of positive acceptance. But the way in which we understand the mode(s) of recognition will typically also make certain other ideas more plausible, and one consequence of understanding institutions in terms of some agreement, collectively held ideas, or joint acts is that it makes sense to think in terms of institutions as sets of rules. And, in turn, if we think that institutions are rules, one very natural way in which rules can be in place is that we collectively accept or agree on them. So these are mutually reinforcing ideas, and they are also found in tandem. For instance, Gilbert (2018, 30) characterizes an institution as “a system of rules that is a blueprint for human behavior,” Guala and Hindriks (2015) have proposed a rules-in-equilibrium account of institutions, and while mainly theorizing institutional facts, Searle (2005, 10) has suggested that “[a]n institution is any system of constitutive rules of the form X counts as Y in C.” A contract theorist like John Rawls (1999, 47-48) takes a similar stance, with an institution being

a public system of rules which defines offices and positions with their right and duties, powers and immunities, and the like. These rules specify certain forms of action as permissible, others as forbidden; and they provide for certain penalties and defenses, and so on, when violations occur.

If we have moved toward doing nonideal social ontology, however, we have also moved away from understanding institutions in terms of something like an agreement or decision, that is, from understanding institutions as if they would have come into place ex nihilo through a singular act of creation, and toward a conception of institutions according to which they are much messier, having evolved over time and building on, and being entangled with, other already existing institutions. On the Rawlsian account, it should however be noted that what institutions ultimately come down to is to establish a certain structured distribution of rights and duties. ¹⁰ Rather than saying that an institution is a set of rules that distributes rights and duties in a certain way, we could instead say that an institution simply is such a distribution. That kind of distribution could come about through a collective agreement on certain rules, but it could also come about through how our micro-level interactions gradually get synchronized and coordinated in ways which mean that different social positions that we occupy come with certain rights and duties. These social positions might vary in character. Some like being a citizen or being a property-holder will be predominately formal and explicit in character, while others, like being a woman or being a man and being white or being a person of color can be informal in character, still coming with certain typical packages of rights and duties, albeit of an implicit kind. Whether formal or informal, explicit or implicit, the relevant rights and duties are about the deontic relations in which we stand to each other: what are we at liberty to do and in which ways are we deontically constrained?

Saying that an institution is a structured distribution of rights and duties means that there is a rule-like character to institutions, but it does not commit us to there being a set of rules that could be fully explicated whenever there is an institution in place. Compared with standard models, this approach involves an ontological shift from understanding institutions in terms of something ultimately abstract (a set of rules) to something ultimately concrete (a set of individuals interrelated in certain ways). This means that if a society has, for instance, an institution of freedom of movement, describing that institution is not about stating an abstract normative rule governing our movements but about describing which moves that are legitimately available moves for which individuals in terms of physically moving about in society. Note that for there to be enough structure for talk of rights and duties to make sense, these moves need not be precisely defined, but our expectations need to be coordinated and synchronized so that there is predictability in terms of what counts as, say, overstepping the relevant boundaries of one’s freedom of movement. People can be relatively dimly aware of why they make the relevant judgments this way or that, and they can rely on similarity assessments rather than applying clear criteria, the important thing for predictability is just that they are in synch. ¹¹

The sense of rights that is of interest here is people’s effective rights, what people can actually do without risking (what is regarded as) legitimate pushback or being sanctioned. One consequence of this view is that while the standard conception opens up the possibility of a gap between having a perfectly egalitarian institution qua a set of rules and there still, in actual practice, being significant inequalities between individuals, this alternative nonideal account closes that gap: to the extent that there are systematic inequalities in actual practice, these are part of the fabric of that institution. Even when there are rules in place in a formal sense, such as certain laws, then rather than being the relevant institutions, these will instead be facilitators of coordination and synchronization, at least to the extent that they feature in people’s governing beliefs. Sometimes they can be very powerful facilitators, but ontologically speaking, the institution qua social object is still just the product of the relevant processes of coordination and synchronization.

If we adopt this alternative conception of institutions and return to thinking about the mode(s) of recognition, it seems reasonable to conceptualize this not in terms of some form of joint embracing of a certain whole, but rather in terms of a vast number of micro-level acts and attitudes reinforcing and maintaining the existing distribution of rights and duties. If we understand the object(s) of recognition in terms of distributions of rights and duties, what primarily characterizes any such distribution will be the ways in which that system enables or disables a variety of different moves. And this is not a question about whether these moves are physically possible to perform or not, but rather about whether they are established as legitimate moves in the relevant contexts. But, one might ask, will relying on a notion of legitimacy not just take us back to casting institutions in a rosy light? No, because by focusing on the legitimacy of specific moves, we are not saying anything about the legitimacy of institutions as wholes (although this is an issue that we will return to in the next section). If we understand the mode(s) of recognition as an embracing of the whole, then that whole arguably gets portrayed as having gotten a stamp of approval, but an institution does not become legitimate as a whole by involving a number of legitimate moves. That certain moves are legitimate is simply a matter of an institution being in place, irrespective of whether that institution as a whole is legitimate or not. In fact, since the focus here is on micro-level moves, participants in the institution need not even really understand the institution as a whole—they can be myopic and simply focus on the moves that are relevant to them.

One reasonable candidate idea for explicating what it means for moves to be legitimate is that they are in accordance with certain kinds of social expectations. When people act in accordance with such expectations we tend to have a sense of them as playing in accordance with the rules of the game (even if we might not have a clear and articulate conception of those rules). Social expectations are typically formed based on previous behavior, so there is no need for anything like an agreement; rather, a gradual process of actions, reactions, and recalibrations of our expectations can have taken us to what is a relatively stable state. Even doing nothing can create such expectations. If you do something to me and I do not protest, you will form an expectation that the next time you do the same thing, I will not protest either. And such expectations would often seem to be normatively laden: since I did not protest, you now think that the action in question is a legitimate move in our relation, in the sense of something that can be done without protest. After all, if I found it objectionable, why did I not protest the first time? When you then make that move again, the threshold for protesting against it has been raised further, it need not be by much, but typically it has been raised. And for every repeated performance, expectations are solidified and that threshold gets a little higher. When you have made that move a great number of times and I still have not objected to it, then at least on a descriptive level, it would seem reasonable to say that in our relation, you can now be understood as having established a right to make that move. This does not mean that our relation cannot be renegotiated, but as things stand now, there is something more in place than a mere behavioral regularity.

If we move up to the level of larger groups, the kind of level where institutions might be said to exist, we will still have expectations on people, although for the most part, these will have a generic character and not the face-to-face quality of the expectations formed in our personal relations. But the most important characteristic of the relevant expectations can still be there, namely, that these are social expectations, where we are aware of each other’s expectations and mutually aware of being aware of our expectations. Since we rely on our expectations to plan our actions, we all tend to, at least on the whole, be invested in aligning with predominant social expectations, as well as policing the behavior of others, so that they too align themselves. This means that once social expectations are in synch, there will be a tendency toward self-reproduction, even though such synched social expectations can certainly still unravel.

The idea here, then, is that if the right kind of social expectations are in place with respect to a certain type of action, it can either count as a legitimate or an illegitimate move. To have the right to perform an action is for that action to count as legitimate move. To have an obligation to perform an action is for any other action to count as an illegitimate move. It is possible that the relevant social expectations are divided or conflicted, so being regarded as a legitimate move cannot merely be a case of not being regarded as illegitimate, and vice versa. Here is a schematic suggestion of how the relevant expectations can be explicated ¹² :

There are a couple of things to note about this explication (apart from it not being fully developed, but more like an example of a kind of account that would fit with the present conception of institutions). One is that the normativity involved here is not based on the reasons for the relevant institution being in place, such as there being an agreement or it being a mutually beneficial solution, but on our reasons for complying given that an institution is already in place. Arguably, this is in line with how we as individuals actually come to enter into institutions, namely, as already populated with people already accepting certain distributions of rights and duties, and where we simply start going along with that state of affairs. This account does not give an answer to the question of how there can be institutions to begin with, but then institutions are understood as concrete entities and for each such entity there will be a concrete and specific process through which it has come into being. Another thing is that since the relation between the deontic relations that hold between us as concrete individuals and the rule-like patterns that are in place is here understood primarily as bottom-up rather than top–down (the latter instead being a characteristic of rules-centric accounts), there being many gray areas where the legitimacy or illegitimacy of certain actions will not be clear-cut is not a sign of an institution being deficient, but is something that is just to be expected since our deontic statuses do not follow from a set of rules (top–down), but emerge from our micro-level interactions (bottom-up).

Finally, while it might perhaps be worried that the sufficiently many relevant others-clause leaves too much open, this is just how it has to be. The reason is precisely that this is an account of institutions that is supposed to capture both strongly hierarchical institutions as well as egalitarian ones. In the latter case, sufficiently many relevant others might simply mean an overwhelming majority, whereas in the former case different people will have different degrees of power, and part of what having such power will be about is to count for more in determining whether the sufficiently many relevant others-clause is satisfied. All the slave-owners will typically be a very small minority, but they will play a vastly disproportionate role in shaping and maintaining an institution of slavery. This kind of account accordingly opens up for how an analysis of power relations always needs to go into our understanding of any real-world institution in terms of how it is grounded in relations between members of the relevant community. Power relations are not just something that is established by institutions being in place but are typically part of the bases that ground their existence.

3. Teleology and Ideology

The above account of institutions relies on a deflated basic notion of legitimacy, where facts about different types of moves counting as legitimate do not say anything about the institution as a whole being legitimate, not even in sense of being perceived as legitimate. Part of the rationale for going in this direction was to steer clear of the tendency to understand institutions in ways that make them come out, at least implicitly, as beneficial or legitimate simply qua being institutions. But to the extent that this is successful, it might perhaps be wondered if this feature of the framework does not leave us with something important left to explain. After all, even though we need to recognize that institutions can often be repressive or oppressive, it would also seem to be the case that people are often quite wholeheartedly on board with institutional arrangements. Sometimes even those who are repressed or oppressed might understand their circumstances as perfectly natural. Indeed, this latter idea has been central to the critical-theory tradition, where notions like ideology, reification, and false consciousness are potential tools in helping us to unmask the contingent and oppressive character of our current circumstances.

A key feature of the present account is that it understands institutions simply in terms of what one might call a deontology being in place, a state of affairs where there is stable and structured sorting of different types of moves in terms of which ones that are allowed and which are not. While the ways in which people possess certain configurations of rights and duties in relation to others will enable and disable a variety of moves, establishing different games that we can play, so to speak, this kind of account leaves unanswered the question of why we are often more deeply committed to playing those games. One possible response here would be to say that institutions should not just be understood in deontic terms, but need to be understood in telic terms as well, that is, in terms of there being a point or purpose to an institution, it serving some function, establishing certain ideals, securing or producing some good(s), and so on. That there is some kind of telic character to institutions is already accepted by some writers; for example, Seumas Miller (2010, 52) has suggested that collective ends being in place is a necessary part of an exchange system being an institution, and in an account like Searle’s deontic powers and status functions come as two sides of the same coin, which means that there is always a definitive function played by our rights and duties being configured in a specific way. ¹³

Given the present account, it would however seem unreasonable to assume this kind of neatness in how things line up. Part of the point of understanding institutions as distributions of rights and duties that grow out of how our micro-level interactions become coordinated and synchronized in various ways is to allow for them to be messy, contested, and rife with tension—structured, but not necessarily well-ordered. Introducing collective goods or definitive functions would simply bring us back to an account where we are assuming that institutions are not just relatively stable, but positively harmonious. For an institution to be stable, or for it to make sense, there need not be a shared conception of certain functions or goals. We can participate in the relevant activities each for our own private reasons. On the conception of institutions proposed here, the structured character of institutions will primarily be understood in terms of different social positions that we occupy and which come with certain typical sets of moves being allowed or disallowed. For a complex of such social positions to form a stable whole, it seems quite reasonable to think that there is a role to be played by widespread ideas about functions served and goods produced by the relevant distributions of rights and duties, but such widespread ideas can be recognized as typically important for explaining the stability of concrete and particular institutions without thereby having to be understood as being essential to something being an institution.

Even if we have a relatively narrow understanding of institutions, we can (and should) still develop a broader social ontology that includes other social phenomena as well—and here telic phenomena can come in. For highly specific practices, it is probably likely that ideas about functions will have a fairly straightforward relation to the moves people are making and how they have synchronized and coordinated their expectations, but such ideas might also exist in relation to larger institutional complexes and then the relation might be more complicated. If we are doing nonideal social ontology, with its emphasis on understanding examples involving dominance and oppression, one important notion to understand would be ideology—especially since the move toward doing nonideal social ontology has partly been motivated here by a concern with social ontology playing an ideological role. And ideology would seem to be something that could be understood in terms of some kind of large-scale teleology providing support to the basic societal distribution of rights and duties. There is, of course, a rich tradition of work on ideology and many different conceptions of ideology in the literature, but for present purposes the objective will simply be to outline a notion of ideology that fits with the account of institutions that has already been provided here, not to argue that a particular understanding of ideology is the best simpliciter (whatever that would mean). This has at least two consequences for which type of account that is reasonable.

To begin with, given that the present approach is based on the negative aim of not playing an ideological role in relation to existing institutions, rather than making a positive emancipatory aim integral to social ontology, it seems reasonable to opt for what Raymond Geuss (1981, 5) has called a descriptive sense of ideology. Such a notion differs from how ideology tends to be understood in the critical-theory tradition, where ideology is typically a pejorative notion, either in a moral or an epistemic sense. ¹⁴ Whichever way we go, ideology will be about maintaining current institutions through different ideas or practices that establish perceptions of systemic legitimacy among participants in relation to the relevant institutions. But given a descriptive account, ideologies can play this stabilizing and legitimizing role even in a just society, and in principle they can even perfectly reflect the actual character of the relevant institutions rather than masking their true workings.

Another consequence is that while the present framework does not necessitate a cognitive conception of ideology, that is, one that understands ideology in terms of beliefs, it can help defuse a worry that one might have with such conceptions. As pointed out by Sally Haslanger (2017a, 150-51), one characteristic of shared beliefs in the social domain is that they can make themselves true. This is a standard tenet in social ontology: that something can be a social or institutional fact, simply because we believe it to be the case. What this possibly means, then, is that if we believe that, say, the definitive function of women is to be caregivers, this would be their function, in a similarly objective sense as bank notes being means of payment. ¹⁵ On the present account, however, basic institutional facts are not about functions, but about rights and duties, and are not established through collective beliefs, but through synchronization and coordination in our micro-level interactions. The beliefs that are part of an ideology will not establish any social or institutional facts, they will only play a role in making it the case that these facts are in place, for example, that women are constrained in certain ways compared with men. ¹⁶

Given the above, we can understand an ideology as a set of widely held beliefs about the function(s) performed by an institution, or a set of institutions, and the reasons for why they are in place, which serves the role of legitimizing the relevant institutions, considered as wholes, in the eyes of those who take part in the activities that these institutions structure. Note that while this type of account does not assume that ideology plays a masking role (in principle we could hold accurate beliefs about functions and reasons), the kind of ideology critique that it enables will be epistemic (and accordingly in line with more traditional critical theory). Precisely because shared beliefs about functions or rationales do not establish any corresponding facts, what this conception opens up is the possibility that there is a discrepancy between (a) the function(s) that people impute to an institution and why they believe it to be in place and (b) the function(s) actually served by the relevant distribution of rights and duties being in place, as well as the actual reason(s) for why that distribution is in place. One important facet of such discrepancies will be about how we as participants might interpret the relevant institutions in place as being natural or inevitable, when they are really contingent and changeable, but perhaps more importantly, such discrepancies will also inevitably raise the question of cui bono?

How common or rare will such discrepancies be? This is not a question that can ultimately be answered through a philosophical analysis, but there is some tentative ground for thinking that such discrepancies will be relatively common, at least to the extent that one finds the following two hypotheses reasonable:

Given these two hypotheses, which at the very least do not appear outlandish, it seems reasonable to assume that even if the conception of ideology proposed here does not mean that ideologies always play a masking or distorting role, they can be expected to typically involve an overstated pretty-picture character. This also means that even a strictly descriptive explication of an institution, together with the ideology that stabilizes it, will potentially have a destabilizing effect, by pointing to the discrepancy between the “official story” about what the institution does and the actual effects of the relevant distribution of rights and duties.

Although there is no room here to develop any such explications, there would be room for Marxist-style analyses of the contrast between, on one hand, the surface-level egalitarian distribution of legal rights and duties characterizing contemporary capitalist societies and, on the other hand, the actual functioning of these institutions in terms of how they stratify those societies. Or one might look at more specific institutions, such as law enforcement in a racialized society, where the privileged majority group might understand the function of the police and the courts simply as maintaining law and order, and be able to make that narrative into the one dominating the political system, but where an important way in which law enforcement in that society actually functions is by suppressing minority groups and segregating members of those groups from society in general. ¹⁸ Or one might look at gender norms and how there might be functionalistic ideas in play about how a certain gendered division of labor is necessary for the reproduction of society, but where reproduction actually takes care of itself, or might perhaps even be hampered by the relevant distribution of rights and duties, ¹⁹ and where what this distribution actually does is primarily to skew the kinds of opportunities that are open to people, depending on their gender, as well as the kinds of outcomes that they tend to achieve.

4. Concluding Remarks

Analytic social ontology has tended to be descriptive in its ambition. And yet, as social ontologists, we are not just bystanders to our own societies and their institutions: when we are articulating analyses of institutions, we are inevitably also participating in the concrete institutions that we inhabit. A worry has been raised here that certain common ideas in social ontology can be understood as potentially playing an ideological role. To the extent that we take this worry seriously, one possible response would be to move to a critical-theory approach instead, accepting that our theories inevitably play a normative role and consciously take on a progressive such role rather than a conservative one. In this article, it has however been argued that it is possible to maintain the descriptive ambition by opting for a nonideal form of social ontology. A specific account has then also been outlined of how institutions can be thus understood and how we can understand ideology and its connection to institutions.

It should be noted, however, that even if the present account is built on an idea of keeping social ontology and political theory conceptually separated from each other, social ontology does have possible implications for political theory. If we seek to assess a society in terms of whether its institutions are just, the way in which we understand institutions will matter for exactly what features of our societies will go into such assessments. One consequence of the present analysis is that such assessments should reasonably focus on patterns in our micro-level interactions rather than on abstract rules or principles, that is, on how we as individuals concretely tend to relate to each other rather than on some assumed underlying blueprint for those relations. Or to put it another way, nonideal social ontology points us in the direction of nonideal political theory.